To run inference on Apple Silicon, pytorch checkpoints have to be exported to format suitable for running on Apple Silicon, detailed instructions and code can be found model_export subfolder. Please ...